源内

Japan's Digital Agency announced the sequential deployment of an AI agent creation environment in its government AI service 'Motochi' and published unique safety standards called 'Motochi Criteria'.

• •The Digital Agency announced on October 20 that it will sequentially introduce an environment allowing government staff to create and run their own AI agents in its generative AI service 'Motochi'.
• •The plan involves deploying the AI agent environment to 180,000 government staff members.
• •The agency also published unique safety rules known as the 'Motochi Criteria' to ensure safe usage.

An article explaining the Digital Agency's plan to provide the generative AI 'Minamoto Uchi' to approximately 180,000 officials across all government agencies, and its efforts regarding confidentiality and cost management.

• •The Digital Agency is providing the generative AI 'Minamoto Uchi' to approximately 180,000 officials across all government agencies.
• •There are challenges to resolve, such as the confidentiality of information and the cost perception of AI operated with tax funds.
• •The article introduces specific AI utilization initiatives considering these challenges.

デジタル庁の生成AI基盤「源内」は、高いセキュリティと運用効率を両立するため、省庁ごとに環境を分離しつつ、インフラ担当者を1〜2人に抑える仕組みを採用している

• •省庁ごとに環境を分離することでセキュリティを確保しつつ、運用負荷の増大を防ぐ設計となっている
• •インフラ担当者をわずか1〜2人体制で抑え、18万人規模の職員が利用する基盤を効率的に運用している
• •この「一見手の込んだ仕組み」により、セキュリティと運用効率化という二律背反を乗り越えている